Ischemic stroke selectively inhibits REM sleep of rats

Abstract

Sleep disorders are important risk factors for stroke; conversely, stroke patients suffer from sleep disturbances including disruptions of non-rapid eye movement (NREM) and rapid eye movement (REM) sleep and a decrease in total sleep. This study was performed to characterize the effect of stroke on sleep architecture of rats using continuous electroencephalography (EEG) and activity monitoring. Rats were implanted with transmitters which enabled continuous real time recording of EEG, electromyography (EMG), and locomotor activity. Baseline recordings were performed prior to induction of either transient middle cerebral artery (MCA) occlusion or sham surgery. Sleep recordings were obtained for 60 h after surgery to identify periods of wakefulness, NREM, and REM sleep before and after stroke. Spectral analysis was performed to assess the effects of stroke on state-dependent EEG. Finally, we quantified the time in wake, NREM, and REM sleep before and after stroke. Delta power, a measure of NREM sleep depth, was increased the day following stroke. At the same time, there was a significant shift in theta rhythms to a lower frequency during REM and wake periods. The awake EEG slowed after stroke over both hemispheres. The EEG of the ischemic hemisphere demonstrated diminished theta power specific to REM in excess of the slowing seen over the contralateral hemisphere. In contrast to rats exposed to sham surgery which had slightly increased total sleep, rats undergoing stroke experienced decreased total sleep. The decrease in total sleep after stroke was the result of dramatic reduction in the amount of REM sleep after ischemia. The suppression of REM after stroke was due to a decrease in the number of REM bouts; the length of the average REM bout did not change. We conclude that after stroke in this experimental model, REM sleep of rats is specifically and profoundly suppressed. Further experiments using this experimental model should be performed to investigate the mechanisms and consequences of REM suppression after stroke.

Figure 1

Sleep EEG recordings on rats after stroke. Stroke was performed on the right side. A representative awake rat’s EEG before (A) and after (B) stroke demonstrates attenuated voltage after stroke over the ischemic cortex. NREM and REM sleep showed no obvious changes over the left (non-ischemic) cortex. EMG demonstrates muscle activity specific to the awake phase.

Figure 2

Spectral analysis of wake and sleep EEG after stroke. This figure depicts only the left cortex, which is contralateral to the stroke. Periods of wake, NREM and REM sleep before and after stroke were identified and then subjected to Fourier analysis. The power spectra for the day prior to surgery was used as a baseline for each individual animal. The baseline was subtracted from power spectra of EEG records performed during the first night after stroke (left column), the first daytime period (center), and second night after stroke (right). The difference spectra were then averaged for all animals. Spectra for sham animals (n=10) are presented in the same panels as animals subjected to stroke (n=8). Only animals that experienced more than 24 NREM or REM epochs per hour were used of spectral analysis. Since there was substantial suppression of REM after stroke, the numbers of animals analyzed for REM after stroke were lower than for other states (n=2 for Day 0 dark and Day 1 light; n=3 for Day 1 dark). One of the sham surgery animals was excluded from REM analysis due to insufficient REM data. Significance is denoted (*p < 0.05 and **p < 0.01).

Figure 3

Temporal evolution of post-stroke NREM delta power. (A) Time course of delta power differences induced by stroke or sham surgery was derived by subtracting baseline delta power (integrated delta power spectrum from specified hourly periods of the circadian cycle on the day prior to surgery) from postsurgical NREM delta power at the same times. Daytime and nighttime periods are shown as open and dark bars, respectively; the first hour 18 represents the onset of the first dark period after stroke or sham surgery. NREM delta power was increased at all time points; differences were significant at times noted (*p<0.03). (B) Circadian variation in NREM delta power in animals prior to stroke or sham operations the day before the operation. These values were used to compute enhancement of NREM delta power after surgery in (A).

Figure 4

Power spectra changes specific to ischemic cortex. A two-step process was used to determine differences in power induced by stroke and specific for the ischemic cortex. First, a similar analysis as in Figure 2 was performed for the right (ischemic) cortex (not shown), resulting in difference spectra that show changes in the ischemic cortex related to stroke. Next, the difference spectra from Figure 2 were subtracted from the difference spectra from generated from the right hemisphere. The resulting analysis is presented according to state (wake, NREM, and REM) at different times after stroke. Significance is denoted (*p < 0.05 and **p < 0.01).

Figure 5

Longitudinal analysis of sleep time after stroke. This analysis shows the effect of stroke or sham surgery on total wake and sleep time compared to presurgery baselines. Grouped analysis of time spent in each state after undergoing right MCA occlusion or sham surgery is displayed. (A) The timeline for the experiment is shown. Dark bars represent nighttime. Sleep was analyzed for at least three days prior to stroke, which was performed on day 0 during the daytime (time window represented by bold bar). After stroke or sham surgery, the animal was returned to monitoring cages; the end of surgery always occurred during the daytime (white bar). Sleep recordings were temporarily halted during surgery; thus, values for sleep during the day on day 0 are not shown. (B) Total, NREM, and REM sleep were recorded over days subsequent to stroke during the day and night. The amount of sleep (see methods) is represented by vertical bars (white bars indicate daytime and dark bars indicate nighttime); days are displayed in the x-axis. Baseline state time was defined as the percentage of time in each state for the day or night prior to stroke. Postsurgery times that differed significantly from baseline state time are denoted by stars and specific p values shown.

Figure 6

Effect of stroke on the number and length of sleep bouts and sleep power. The average length (A) of REM sleep bouts on the day after stroke or control procedure is displayed. (B) The number of REM bouts as a function of bout length (minutes) is shown per 24 hour period. The control values (white boxes) represent daily numbers of REM bouts prior to surgery. The numbers of bouts of REM after stroke were significantly decreased (red boxes; for 0.2-1.0 minute bouts p<0.02; for 1.2-2.0 minute bouts p<0.02); sham surgery did not change the number of REM bouts.